By Phillip Mboya
June 2012
Most countries in the Macroeconomic and Financial Management Institute of Eastern and Southern Africa (MEFMI) region have decided to conduct foreign private capital (FPC) surveys in order to, among others, monitor their magnitude, source country, sectoral preference and their impact to the host economy in terms of employment and revenue generation, market linkage and technology transfer. In the monitoring process, countries use different survey methodologies which make comparison and sharing of FPC statistics across countries a great challenge. Using the case of Tanzania the study attempts to identify weaknesses in the methodology for compiling FPC statistics in the MEFMI region.

The study finds that maintaining updated register is still a challenge due to existence of multi sources of updating the sampling frame. Consequently, huge amount of resources is used for locating enterprises that either do not exist or have changed status such that they do not qualify for survey. In this respect, continuous update of the frame is of paramount importance.

The study finds that the use of appropriate sampling and estimation techniques is also a challenge. For instance, Tanzania uses purposive sampling whereby the choice of sample size is based on the resource envelops. Sample selection involves setting a threshold level in which case enterprises laying below the threshold are automatically left out. Information obtained from samples constructed in this manner can hardly be used to estimate properties of the population. In addition, estimation procedures used to estimate population figures rest on assumptions that are difficult to justify under the existing statistical theories. The study proposes the use of probability sampling, particularly simple random and stratified sampling because the two methods are simple to implement, allow researchers to draw valid inferences about the entire population, they are free from classification error, require minimum prior knowledge about the population and lead to a high level of precision.